Redistribution Reactions in the Halides of Carbon, Silicon, Germanium

Redistribution Reactions in the Halides of Carbon, Silicon, Germanium and Tin. George S. Forbes, Herbert H. Anderson. J. Am. Chem. Soc. , 1944, 66 (6)...
0 downloads 0 Views 375KB Size
June, 1944

REDISTRIBUTION REACTIONS OF SOME HALIDES [CONTRIBUTION FROM

THE

93 1

CHEMICAL LABORATORY OF HARVARD UNIVERSITY ]

Redistribution Reactions in the Halides of Carbon,Silicon, Germanium and Tin BY GEORGE S. FORBES AND HERBERT H. ANDERSON The fundamental papers of Urey and co-workers on isotopic exchange equilibrialJJ emphasize the fact that not even units so closely similar as isotopic atoms attain a truly randoin distribution among similar molecules. These investigators have set up general equations for calculation of equilibrium constants based on the distribution functions of diatomic and polyatomic molecules, and have quantitatively predicted those important deviations from random distribution which have served them so well in their separations of isotopes. One might expect a firwri that any exchange reactions involving different organic groups, or even different halogen atoms, would exhibit substantial deviations from simple probability laws. An important series of papers by Calingaert and ~ o - w o r k e r sreports, ~ ~ ~ ~however, ~ ~ ~ ~ ~that ~~ random exhange of organic radicals occurs within experimental error among certain alkyls, chloroalkyls and esters, likewise different halogen atoms among ethylene halides. From the laws of probability alone they work out a general equation for quantitative prediction of equilibria in such reactions, and stress the absence of chemical or energy factors affecting random distribution. They emphasize that “the redistribution reaction is adaptable as a new method of organic synthesis”.1° It would seem that inorganic chemists have long possessed at least an instinctive realization of the possibilities of the redistribution reaction, especially in the synthesis of mixed halides containing a single atom of a quadrivalent element. For instance, Bessonl’ obtained all the chloroiodides of silicon by passing iodine monochloride over silicon heated to redness. He found that if a silicon chloroiodide (plus iodine) was passed over heated silicon, silicon tetrachloride was one this of the products, and he astutely remarks ‘ I . is not surprising if one takes account of the partial dissociation of chloroiodides under the action of heat.” Subsequently12 he obtained mixed halides froiii the action of hydrogen bromide on phosphoryl chloride. He points out that the boiling point of POClBr2, when distilled at normal pressure, is not fixed, since a t this temperature (about 165’) it undergoes a progressive decomposition according to the equations

..

++

2POCIBrz = POClrBr POBrs and 3POClBrt = POCIS 2POBra ~~~

(1) Urey and Rittenberg, J . Chcm. Phys., 1, 137 (1933). (2) Rittenberg, Bleakney and Urey, ibid., 9, 48 (1934). (3) Urey and Greiff, THIS JOURNAL, 67; 321 (1035). (4) t o (7) ibid., 61, 2748, 2755, 2758, 3300 (1939). ( 8 ) and (9) ibid., BS, 1099, 1104 (1940). (10) Reference 4, p. 2754. (11) W n , Compl. raitd., ill,1314 (1R91). (12) Beason, ibid., 199, 814 (1890).

Lecompte, Volkringer and Tchakirian118 improving a procedure of Besson,14 heated chloroforni with bromine in a sealed tube a t successively higher temperatures up to 250°, allowing hydrogen bromide to escape after each heating. By fractional freezing, in addition to distillation, all the mixed halides were obtained. Raederlb measured molar elevations of boiling points and molar depressions of freezing points in a great number of solutions made by dissolving a given halide of a given element from a middle group of the periodic system in excess of a second halide in the same category. By plotting kAT/mole against mole per cent. of the first halide, he arrived a t the following conclusions. The halides of phosphorus, arsenic, antimony, titanium, tin (and by prediction germanium as well) easily exchange their halogen atoms so as to enter into readily mobile equilibria and form mixed halides incapable of isolation in the solid or in the gaseous phase. On the other hand, the tetrahalides of carbon and silicon, which are characterized by great chemical inertness, form mixed halides only with difficulty and only under special conditions. Once formed, however, they can be isolated as pure compounds. Raeder cites Triimpy’s analyses of Raman spectra of such systemsl6 as confirmation of the above interpretations. Papers of references 1 to 14 do not contain quantitative data on yields, so that closeness of approach to random distribution cannot be figured. Our interest in the stabilities of halides, cyanates and thiocyanates prompted a survey of the redistribution problem over Group IVA of the periodic system, in which a t least the mixed chlorides of four elements enjoy some measure of stability. A subsequent paper by Anderson1’ describes the preparation and properties of three new compounds, SiClaNCO, SiCt(NC0)2 and SiCI(NC0)a and shows that these, as well as mixtures of the end members of the series, attain at elevated temperatures distributions random within experinirntal error. In the present paper it is shown that their behavior fits into a larger scheme of relative stabilities. Carbon Ch1orobromides.-No reaction between carbon tetrachloride and tetrabromide was detected after seven hours in a sealed tube a t 170°, or in a second experiment in which dry aluminum chloride was added. In the three (13) Lecornpte, Vqlkringer and Tchakirian, ibid., 904, 1927 (1937). (14) Besron, $bid., 114, 222 (1892). 115) Raeder, 2. w i w p . dlgem. Chcnt., %IO, 145 (1933): from Det. K g l . Norske Vidrnrk Sslrk. Skriflsr, 8 , 1 (1929). (10) Trtlmpy. Z. Physik, 66, 780 (ISSO); 68, 676 (1931). (17) Anderaon, THISJOURNAL, 66, 834 (1944).

GEORGE S. FORBES AND HERBERT H. ANDERSON

932

Vol. 66

TABLB I Compounds

3 CCb,CBrc 4 CCL, CBrd 5 CCbBr, CClZBn

h D I W R I B U T I O N OF THE CARBON c H L O R O 3 R O ~ D B S ; Mole, % C1 %cch CCliBr

0.723

o.623 0.67

i

obs. 29.6 ~ a l c27.3 .~ robs. 12.9 calc.' 15.1

{

2

.

4

g.2

38.6 41.9 38.8 36.5 41 39.7

CATALYST

CCl:Bri

CClBr8

CBrr

23.5. 24.1 34.5 33.1

6.8 6.1 12.1 13.3 11 9.6

1.5 * 0.6 1.7 2.0 l h 1.2

28 29.3

f

TABLE I1 REDISTRIBUTION OF THE SILICON CFXLOROBROMIDBS; No CATALYST Compounds

Mole. $6 C1

1 Sick, SiBr4

0.50

2 SiCltBr

0.75

3 SiClrBr, SiClBrt

0.60

% Sic14 (obs. 15.3 calc. 6.25 obs. 31.0 calc. 31.6 obs. 12.1 calc. 12.96

experiments listed in Table I, slightly moistened aluminum chloride was used as a catalyst, and the pressure of hydrogen chloride in the sealed tube was apparent. Random distribution was now attained within experimental error. The calculated mole percentages in Table I are figured from Calingaert's equations.* We have found no previous record of a direct reaction between carbon tetrachloride and carbon tetrabromide. Experimental.-Carbon tetrabromide was precipitated from a water solution containing 25 g. of sodium hydroxide, 1 cc. of acetone and 5 cc. of bromine per liter as recommended by Wallach.'8 About 120 g. of crude product was recovered, washed, dfied by suction and fused. A mixture with carbon tetrachloride in a known mole ratio was heated in a sealed tube a t 170' for seven hours. Decomposition of carbon tetrabromide would have been excessive a t higher temperatures or over longer periods. The reaction, if any, was less than 10% of the whole. A second experiment, with introduction of dry aluminum chloride, generated no pressure in the tube and appeared to yield no chlorobromides. In the succeeding experiments the aluminum chloride was spread out on paper, and breathed upon to activate it, presumably through formation of hydrogen chloride. Redistribution now occurred readily. After washing with caustic solution and drying with calcium chloride, the liquid mixture, 5-7 ml. in volume, was distilled using a column 30 cm. high and 7 mm. in diameter with a spiral of No. 23 copper wire having 80 turns, and was surrounded with a vapor jacket. Simultaneous readings were made of temperature, and of the depth of liquid collected in small flat-bottomed tubes backed by a millimeter scale. Satisfactory "flats" were observed at the boiling points of the first four members of the series, namely, 77', 104', 134O, and 160'. The cuts were made half-way between these temperatures. Small corrections were applied t o the fractions for the estimated molar volumes of the respective chlorobromides. We encountered difficulties in separating the last of the chlorotribromide from tetrabromide in the last, very small, fraction. Our system was further complicated by presence of the compounds GBrr and GBrs. In Experiment 3, presence of carbon tetrabromide was established qualitatively, but the figures in the last column indicate nothing beyond orders of magnitude. The other volumes, read within 0.05 ml., yielded mole percentages significant, we believe, within 270. Silicon Chlorobromides.--Silicon tetrachloride and tetrabromide were heated for seventy hours in a sealed (18) Wallach. A n n . , 171, 149 (1893).

SiCtBr

SiClrBri

SiClBrf

SiBrc

22.5 25.0 39.8 42.2 38.8 34.56

25.1 37.5 26.5 21.1 28.4 34.56

22.5 25.0

14.6) 6.25 (0) 0.4 1.8 a 2.66

2.7 4.7 19.0 15.36

tube a t 140" without a catalyst. The physical properties of the fractionated products were measured, or remeasured to improve previously existing data. The figures are being submitted for separate publication in THIS JOURNAL. For redistribution studies, however, the different reactants in known mole ratios were passed through a tube at 600' approximately. Equilibrium was established quite rapidly, but the first experiment was evidently too hurried, so that mixtures 2 and 3 were passed a t a half and a quarter of this rate, respectively. Table I1 summarizes the results. Corrections for the respective molar volumes have been applied. Experimental.-In Experiment 1.12 ml. of mixed liquids passed through a tube of 17-ml. volume a t about 600' in forty-five minutes, but about 20% of the reactants escaped reaction. Table I1 shows that the second and third rates were not too rapid at the temperatures prevailing. During distillation in the apparatus described above, the temperature moved very slowly, over considerable volume intervals, in the vicinity of the boiling points of the five pure compounds, 57', SO', 104.5', 128' and 153', respectively. The last half of fraction 4 in Experiment 3 boiled at an average temperature of 132' instead of 128O, indicating the presence of some 16% of silicon tetrabromide. The volumes of tetrabromide in Experiments 2 and 3 indicate orders of magnitude only. Since the members of the series carbon tetrachloride through carbon tetrabromide require a catalyst for rearrangement at 170°, while the members of the corresponding silicon series rearrange without a catalyst, it appears that carbon tetrahalides are much more stable with respect to rearrangement than those of silicon. Silicon Ch1oroiodides.-Treatment of silicon tetrachloride with fused potassium iodide yielded a relatively small amount of SiCLI. Upon passage through a hot tube, a mixture of all five members of the series was obtained. At 58', 114' and 172', the boiling points of the first three members, the "flats" were especially sharp. The fractions in Table I11 approximate random distribution.

TABLE I11 Mole % observed ' calMole % culated

Sic14

SiClaI

SiCldl

Sic118

Si14

30.4

42.7

22.0

3.7

1.2

31.6

42.2

21.1

4.7

0.4

Silicon chlorobromides, chloroiodides and chlorocyanates attain random distribution without a catalyst. If fractionation of such mixtures is complete within a few hours,

F~DISTRIBUTION REACTIONS OF SOME HALIDES

June, 1944

933

the products are separated without appreciable rearrangement. 210 Germanium Chlorobromides.-Booth and Morrislo prepared the closely related germanium fluorochlorides. Held at -78', these rearranged to the extent of 60% within two weeks. We passed an equimolar mixture of tetrachloride and tetrabromide, prepared separately from 190 the elements, through a hot tube a t 600'. Five milliliters of the liquid product was distilled through the column described above, but more rapidly than the carbon and silicon chlorides. The initial rise in the actual distillation 170 curve (Fig. 1) is consistent with a small percentage of tetrachloride and a low reflux ratio. Next, GeCI, and GeCbBr are formed a t the expense of higher chlorobromides and bled off. The first intermediate, GeClaBr, evidently distills 150 off at 112O,between 25 and 36 atom per cent. of bromine U allowing for products already distillcd. Its calculate! kl boiling point, log', is one-quarter of the way between 83 i% and 186.5', the boiling points of the end members. On the B 130 whole the separate existence' of germanium trichlorobromide, boiling point 111 * 3 , is supported by the data; Redistillation of 12 ml. of distillate collected between 96 and 122' yielded relatively little a t log', and large fractions approximating the end members. Evidently re110 arrangement is quite rapid between these temperatures. In another distillation, still more rapid, inconclusive retardations of the temperature rise seemed to occur near 135' and 161°, but we are reluctant to offer these in evi90 dence. The upper half of the actual distillation curve shown in Fig. 1 resembles a curve for a mixture of stable end members. 10 30 50 70 90 Mixed Stannic Halides.-Bessonm found it impossible to separate stannic chlorobromides from each other by Volume per cent. repeated fractionation even under greatly diminished Fig. I.-Distillation curves of germanium ( 8 )and pressure. He states that this would result in total destannic (0) chlorobromides. composition of the desired products into stannic chloride and stannic bromide. By fractional freezing he then obtained three products which he held to be the three halogen atoms are distributed among the five chlorobromides of tin-quite nearly pure. members of the series according to the laws of Raeder's freezing point curves for the systems SnBr, SnIrzl and SnCL SnBr422indicated continuous series of probability The corresponding rearrangements in the silimixed crystals without compound formation. He showed*' why Besson's procedure led to fractions having con series (chlorobromides and chloroiodides) the properties observed by him. I n the system SnC14 in seven hours at 140' without catalyst are parSI&, however, he found evidence of the compound SnC121222 having an incongruent melting point, and stable only be- tial, but complete after passage through a tube a t 600'. These random mixtures can be fully low - 4 7 O . As we were disinclined t o undertake a further study separated by fractional distillation a t normal of this problem, we merely passed equal volumes of stannic pressure without appreciable rearrangement. chloride and bromide through the hot tube, and redistilled Distillation curves of the hot tube product the condensate in the same apparatus as used with the other chlorobromides, over a half-hour period. The dis- from germanium tetrachloride and tetrabromide tillation curve (Fig. 1) shows no retardations of tempera- show definite indications of the new compound ture increase a t any of the calculated boiling points of the GeC13Br a t 112' approximately. Upon redischlorobromides. Its initial slope suggests rapid replacement of tetrachloride a t the expense of chlorobromides. tillation of such a fraction, rapid redistribution It lacks the steep middle portion to be expected, in this takes place. apparatus, of two non-reacting liquids boiling a t 114' and The slope of a distillation curve of the stannic 202', respectively. Extrapolating the findings for carbon, chlorobrornides showed no irregularities near the silicon and germanium, one would predict a random mixture of all members of the tin series, very rapidly boiling points calculated for the mixed comrearranging a t all the distillation temperatures, in harmony pounds. Some indication of redistribution was with Raeder's conclusions. It would be of interest to obtained, although the individual compounds determine whether mixed tin halides containing one or could not be isolated, presumably because of more fluorine atoms could be isolated. redistribution during distillation.

9 $

g

+

+

+

Summary Mixtures of carbon tetrachloride and tetrabromide, as well as the carbon chlorobromides, rearrange in seven hours at 170' with slightly moistened aluminum chloride as a catalyst, until (19) (20) (21) (22)

Booth and Morris, T H I S JOURNAL. 68, 90 11936). Besson, Cotnbf r e n d , 124, 683 (1897). Rneder, Z anorg. aflgcm Ckem 130, 325 (1923) Rneder, ibid , 162, 223 (le27).

None of the phenomena observed were inconsistent with the hypothesis that all possible members of the germanium and tin chlorobromide series coexisted in random distribution. The stability, against rearrangement, of the chlorobromides of Group IVA decreases steadily from carbon to tin. CAMBRIDGE, MASS.

RECEIVED J A N U ~ ~ R 24, Y 1944